Number of species as estimated by the continuum counting method described in the text, *S*_{C}(*A*), for self-similar (*Left*), compact cluster (*Center*), and Gaussian cluster distributions (*Right*) of individuals, plotted as a function of sampling area, *A*. *Insets* show examples of the individual distributions for a randomly chosen species. The same cover distribution *p*(*c*) as for the power-law correlation function case was used for the compact and Gaussian case. The total cover for each species was divided in compact groups and randomly spread over the area. The Gaussian case was identically managed, except that individuals in each cluster were distributed by following a Gaussian distribution (plots shown have a clustering coefficient of μ = 1.5 and use the same abundance distributions as the compact cluster case). In all three cases basically the same SARs are obtained, showing that differences in cluster structure only produce lower-order corrections to the SAR. All SARs display a linear correlation coefficient *r*_{l}^{2} > 0.99.

## PubMed Commons